[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] Science Fair Project

On 11/16/12 10:49 AM, mrapol@xxxxxxxxxxxx wrote:
One way to vary the output by focusing on one component might be with
the spark gap:

simple spark gap
multiple spark gap
quenched spark gap
magnetic quench
'hyperbaric' gap with air blown through or across it, etc.


This is a good experiment, but has some problems in the science fair context. And it depends on what level fair you're competing at.. 8th grade school vs county, state, or ISEF, because the judges are qualitatively different and look for different things..

I'm going to comment here on the assumption that the ultimate judging is at the state or big county fair level or bigger, so if you're not in that class, then you can ignore a lot of what I have to say.

The issues with this have to do with the hypothesis and quantitative aspects.. That is, a "cut and try" or "we tried 6 different things and #3 worked best" doesn't do well, because it's not science nor engineering: it's random tinkering. That is, you want to avoid getting the result (e.g. quenched gap works best) and not know *why* it works best, or if it might have worked best because of some other factor.

To be a winner, there has to be an element of applying theory to predict behavior, and experiment to see if it matches theory. And ideally it needs to be quantative, preferably on a continuous basis (that's just because of bias on the part of judges that prefer it)

Remember, science fairs are about winning the fair, not necessarily advancing the state of knowledge, so there's some peculiarities of the process that need to be addressed.

OK.. to the specifics of why this particular experiment would be hard to do well (although it *can* be done, and done well)

1) you'd have to make measurements on the gap by itself to determine the electrical properties. For instance, you'd need to measure the voltage drops and currents.

2) you have to have a model of tesla coil performance, where you plug in the measured (or predicted) properties of the gap, and make a *prediction* of the performance of the coil.

This is the *key* to the whole hypothesis thing.. you have to have a predicted outcome based upon theory and prior measurements... then you go do the experiment and see how closely theory and experiment match. Then you do some statistical analysis to see if the difference is *significant*.

If you don't do this, then your project is a "cut and try", and no different from the "does rock or classical music make plants grow better" kind of thing.

Now, if you can come up with some way in which the various gaps affect the performance: maybe their quenching behavior is different, and you have quantitative theory as to why quenching affects spark length. Or, maybe it's just about loss. You feed 200 watts into the coil system, and gap 1 has 50 watts loss and gap 2 has 100, and lo and behold sparks are longer with gap 1.. But to make it good, what you should be able to do is reduce the input power with gap 1 down to 150 Watts (same 100 watts after gap losses) and get identical power output.

There is also a problem with Tesla Coil projects in general. You're being judged against other projects in your category. It might be "Engineering" so you're up against everything from wind turbines to robots to solar power collectors, etc. So you need to distinguish your project from the other neat and nifty things.. the way you do that is with theory and math and statistics.

What the judge really, really wants to see is that you did some research ahead of time, and were able to make a quantitative prediction of outcome based on theory.

Even worse is if you are being judged up against more "classical hard science" kinds of projects.. biology or medicine, where they are looking at comparing the effects of varying concentrations of a substance on something growing in a petri dish. The engineeringy judges have no problem judging the biology project: they look at experimental methods and statistics. The biology judges, on the other hand, have a tough time fairly judging engineering projects: they're looking for things like "control groups" and "null hypothesis" and "Chi-square analysis of significance"

Science judges also like to see experiments where there is a "dose response" effect. That is, rather than discrete effects from something being there or not, they like to see a continuum: setting the knob to zero does A, Setting it to 1 does A+B, setting it to 2 does A+2B, etc. (or square law or something predictable)

So it's a challenge to come up with a good project using a tesla coil where "performance of the coil" is the variable.

Now... one thing that might be good is to use the Tesla coil as a part of an experiment. Maybe you are developing a way to measure the electric and magnetic fields of a HV source? Could you figure out whether there is really an "ion cloud" around the top load?

What about studying spark growth? Some video cameras these days have a high speed frame rate option (slo-mo), and you could see how the spark grows on successive "bangs". the theory is a bit complex, but it's doable.

Tesla mailing list